pfSense laggy after WAN outage
-
After a rare long run of no WAN outages, it happened... power brownout knocked the ISP offline for a few hours. When that was repaired, the connection came up but I struggled to access the pfSense GUI. Pings to Google DNS were intermittent, as well. Eventually, rebooting the firewall resolved the issue.
Anyone experience this after a WAN outage?
-
Check the system logs for the period before the reboot. Was there something stuck in a loop perhaps?
-
This appears to be when the link came back up:
<11>1 2026-05-24T12:33:35.817649-04:00 BnK.local.lan check_reload_status 497 - - Could not connect to /var/run/php-fpm.socket
<11>1 2026-05-24T12:33:36.819648-04:00 BnK.local.lan check_reload_status 497 - - Could not connect to /var/run/php-fpm.socket
<13>1 2026-05-24T12:33:36.828676-04:00 BnK.local.lan check_reload_status 497 - - Linkup starting re1
<27>1 2026-05-24T12:33:36.930634-04:00 BnK.local.lan rtsold 61214 - - <cap_rssend> sendmsg on re1: Permission denied
<6>1 2026-05-24T12:33:36.939622-04:00 BnK.local.lan kernel - - - re1: link state changed to UP
<171>1 2026-05-24T12:33:37.000000-04:00 BnK.local.lan nginx - - - 2026/05/24 12:33:37 [error] 91972#100209: *5521 connect() to unix:/var/run/php-fpm.socket failed (61: Connection refused) while connecting to upstream, client: 192.168.1.5, server: , request: "POST /widgets/widgets/disks.widget.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php-fpm.socket:", host: "192.168.10.2", referrer: "http://192.168.10.2/"
<171>1 2026-05-24T12:33:37.000000-04:00 BnK.local.lan nginx - - - 2026/05/24 12:33:37 [error] 91972#100209: *5521 connect() to unix:/var/run/php-fpm.socket failed (61: Connection refused) while connecting to upstream, client: 192.168.1.5, server: , request: "POST /getstats.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php-fpm.socket:", host: "192.168.10.2", referrer: "http://192.168.10.2/" -
Then a little later in the logs:
<11>1 2026-05-24T12:37:59.387615-04:00 BnK.local.lan check_reload_status 497 - - Could not connect to /var/run/php-fpm.socket
<11>1 2026-05-24T12:37:59.432613-04:00 BnK.local.lan check_reload_status 497 - - Could not connect to /var/run/php-fpm.socket
<13>1 2026-05-24T12:37:59.950848-04:00 BnK.local.lan check_reload_status 497 - - updating dyndns wan
<27>1 2026-05-24T12:37:59.980438-04:00 BnK.local.lan php-fpm 94435 - - /rc.linkup: Hotplug event detected for WAN(wan) dynamic IP address (4: dhcp, 6: dhcp6)
<27>1 2026-05-24T12:37:59.980484-04:00 BnK.local.lan php-fpm 94435 - - /rc.linkup: DEVD Ethernet detached event for wan
<11>1 2026-05-24T12:37:59.989717-04:00 BnK.local.lan check_reload_status 497 - - Could not connect to /var/run/php-fpm.socket -
-
@Gertjan Thank you... I will do some digging. I assume that if the PHP is unavailable, this is why the GUI was so very unresponsive.
-
Yes without any available PHP processes you will see the login screen but nothing further in the webgui.
The WAN is on re1 there I assume?
-
@stephenw10 Correct... re1 is the WAN interface.
Odd though... I could navigate enough to get it rebooted without going to the closet, albeit Very Slowly. -
Hmm, well it must have had some processing still available then. At least one PHP process not completely locked. Unless you were using SSH which doesn't require php.
-
@itsbry didn't see you mention what version of pfSense? That might be relevant.
-
@stephenw10 Not SSH...
@luckman212 Apologies!
2.8.1-RELEASE (amd64)
built on Tue Sep 9 12:29:00 EDT 2025
FreeBSD 15.0-CURRENT -
Hmm, well if the link is flapping continuously it could simply be the php scripts that are run each time piling up.
-
We had another outage, this one due to power... remote access stayed down and the GUI was largely unresponsive until the pfSense box was rebooted. While I would love to solve the laggy GUI problem, I'm more concerned about the actual WAN link also being unavailable.
Below is a repeating section from the System/General logs:
2026-06-01 17:24:38.286730-04:00 check_reload_status 507 Reloading filter
2026-06-01 17:24:38.286714-04:00 check_reload_status 507 Restarting OpenVPN tunnels/interfaces
2026-06-01 17:24:38.286694-04:00 check_reload_status 507 Restarting IPsec tunnels
2026-06-01 17:24:38.286604-04:00 check_reload_status 507 updating dyndns WAN_DHCP
2026-06-01 17:24:37.201132-04:00 php-fpm 52657 /rc.dyndns.update: phpDynDNS (): No change in my IP address and/or 1 days has not passed. Not updating dynamic DNS entry.
2026-06-01 17:24:37.201105-04:00 php-fpm 52657 /rc.dyndns.update: Dynamic Dns (): Current WAN IP: 64.179.196.79 Cached IP: 64.179.196.79
2026-06-01 17:24:37.200708-04:00 php-fpm 52657 /rc.dyndns.update: Dynamic DNS custom (): _detectChange() starting.
2026-06-01 17:24:37.199924-04:00 php-fpm 52657 /rc.dyndns.update: Dynamic DNS: updatedns() starting
2026-06-01 17:24:36.224763-04:00 php-fpm 14769 /rc.linkup: DEVD Ethernet detached event for wan
2026-06-01 17:24:36.224738-04:00 php-fpm 14769 /rc.linkup: Hotplug event detected for WAN(wan) dynamic IP address (4: dhcp, 6: dhcp6)
2026-06-01 17:24:36.224515-04:00 check_reload_status 507 Reloading filter
2026-06-01 17:24:36.186232-04:00 check_reload_status 507 updating dyndns wan
2026-06-01 17:24:36.157989-04:00 rtsold 38320 <cap_rssend> sendmsg on re1: Permission denied
2026-06-01 17:24:35.412931-04:00 check_reload_status 507 Restarting IPsec tunnels
2026-06-01 17:24:35.053763-04:00 php-fpm 32861 /rc.newwanip: IP Address has changed, killing all states (ip_change_kill_states is set).
2026-06-01 17:24:34.385613-04:00 php-fpm 32861 /rc.newwanip: rc.newwanip: on (IP address: 64.179.196.79) (interface: WAN[wan]) (real interface: re1).
2026-06-01 17:24:34.385423-04:00 php-fpm 32861 /rc.newwanip: rc.newwanip: Info: starting on re1.
2026-06-01 17:24:33.388013-04:00 php-fpm 29473 /rc.linkup: Starting rtsold process on wan(re1)
2026-06-01 17:24:33.387895-04:00 php-fpm 29473 /rc.linkup: Starting DHCP6 client for interfaces re1
2026-06-01 17:24:33.376306-04:00 php-fpm 29473 /rc.linkup: Accept router advertisements on interface re1
2026-06-01 17:24:33.376236-04:00 php-fpm 29473 /rc.linkup: calling interface_dhcpv6_configure.
2026-06-01 17:24:33.375537-04:00 check_reload_status 507 rc.newwanip starting re1
2026-06-01 17:24:33.359048-04:00 kernel - re1: link state changed to UP
2026-06-01 17:24:33.284980-04:00 check_reload_status 507 Linkup starting re1
2026-06-01 17:24:29.339108-04:00 kernel - re1: link state changed to DOWN
2026-06-01 17:24:29.188463-04:00 check_reload_status 507 Linkup starting re1
2026-06-01 17:24:29.158100-04:00 php-fpm 29473 /rc.linkup: HOTPLUG: Configuring interface wan
2026-06-01 17:24:29.158084-04:00 php-fpm 29473 /rc.linkup: DEVD Ethernet attached event for wan
2026-06-01 17:24:29.158016-04:00 php-fpm 29473 /rc.linkup: Hotplug event detected for WAN(wan) dynamic IP address (4: dhcp, 6: dhcp6)
2026-06-01 17:24:29.157850-04:00 check_reload_status 507 Reloading filter
2026-06-01 17:24:29.133470-04:00 php-fpm 4891 /rc.linkup: The command '/usr/local/sbin/unbound -c /var/unbound/unbound.conf' returned exit code '1', the output was '[1780349069] unbound[4656:0] error: bind: address already in use [1780349069] unbound[4656:0] fatal error: could not open ports'
2026-06-01 17:24:09.564211-04:00 tailscale 97995 Unable to find device tailscale0
2026-06-01 17:24:04.699348-04:00 rtsold 82630 <cap_rssend> sendmsg on re1: Permission denied
2026-06-01 17:24:03.489767-04:00 tailscale 90301 Waiting for device tailscale0
2026-06-01 17:24:03.469419-04:00 check_reload_status 507 Reloading filter
2026-06-01 17:24:03.177268-04:00 php-fpm 32861 /rc.start_packages: Restarting/Starting all packages.
2026-06-01 17:24:02.171606-04:00 check_reload_status 507 Reloading filter
2026-06-01 17:24:02.171505-04:00 check_reload_status 507 Starting packages
2026-06-01 17:24:02.171393-04:00 php-fpm 3392 /rc.newwanip: pfSense package system has detected an IP change or dynamic WAN reconnection - 0.0.0.0 -> 64.179.196.79 - Restarting packages.
2026-06-01 17:24:00.679278-04:00 rtsold 82630 <cap_rssend> sendmsg on re1: Permission denied
2026-06-01 17:24:00.157050-04:00 php-fpm 3392 /rc.newwanip: Creating rrd update script
2026-06-01 17:24:00.152010-04:00 php-fpm 3392 /rc.newwanip: Resyncing OpenVPN instances for interface WAN.
2026-06-01 17:23:59.704187-04:00 php-fpm 32861 /rc.dyndns.update: Dynamic DNS () There was an error trying to determine the public IP for interface - wan (re1 ).
2026-06-01 17:23:59.703493-04:00 php-fpm 32861 /rc.dyndns.update: Dynamic DNS: updatedns() starting
2026-06-01 17:23:59.149261-04:00 php-fpm 3392 /rc.newwanip: Dynamic DNS () There was an error trying to determine the public IP for interface - wan (re1 ).
2026-06-01 17:23:59.148606-04:00 php-fpm 3392 /rc.newwanip: Dynamic DNS: updatedns() starting
2026-06-01 17:23:58.689506-04:00 check_reload_status 507 Reloading filter
2026-06-01 17:23:58.689447-04:00 check_reload_status 507 Restarting OpenVPN tunnels/interfaces
2026-06-01 17:23:58.689374-04:00 check_reload_status 507 Restarting IPsec tunnels
2026-06-01 17:23:58.689201-04:00 check_reload_status 507 updating dyndns WAN_DHCP
2026-06-01 17:23:58.687385-04:00 rc.gateway_alarm 10170 >>> Gateway alarm: WAN_DHCP (Addr:64.179.196.1 Alarm:down RTT:0ms RTTsd:0ms Loss:100%) -
Hmm, the log shows the WAN reconecting and correctly pulling a new lease etc.
But it physically lost link and reconnected. What is the NIC actually connected to? Did you have a modem reboot?
-
@stephenw10
Power was down for only 10-15 min, so the batteries held the rack up... I do not believe the modem power-cycled during this time (I do not have log access to that device, ISP locked). Layout is below:
Modem -> pfSense -> Cisco3850-L3 -> LANThe firewall does monitor the WAN link but takes no action on it, assumes it is Up all the time.
-
@itsbry said in pfSense laggy after WAN outage:
The firewall does monitor the WAN link but takes no action on it, assumes it is Up all the time.
That's important info.
This means that messages like "kernel - re1: link state changed to DOWN (or UP)" are not initiated on the pfSense side, but on the 'other' side = the modem.
Modem can do so (IRC) to 'signal the downstream device that the renegotiated an upstream 'ISP' connection. Bit ... it's a long time I've seen a modem type device.Btw : the power down was locally (your place/office) or more widespread ? If it's the area or bigger : ISP can also have issues when their stuff goes off line for a while. They will, however, never admit this, so people tend to search (and destroy while doing so) the issue locally, while the real issue is 'upstream'. We've all seen this before.
If it's not the DNS, it's the f**ing ISP ^^ -
@Gertjan
This was absolutely a widespread outage... heavy rain and wind were passing through the area, downing trees etc. I'm surprised that it was only down for a short time.
The ISP is still only providing a coaxial connection and they don't allow modem access so you're right, we'll only guess what is going on upstream. I will say though... the GUI and WAN link were surely to remain in the impaired state until reboot. I wonder if there is something that could trigger a soft reset or something when this happens?For clarity, I only rebooted the pfSense, nothing else. I have confirmed that all devices in the rack stayed up during the event.
-
@itsbry said in pfSense laggy after WAN outage:
2026-06-01 17:24:33.359048-04:00 kernel - re1: link state changed to UP
2026-06-01 17:24:33.284980-04:00 check_reload_status 507 Linkup starting re1
2026-06-01 17:24:29.339108-04:00 kernel - re1: link state changed to DOWN
2026-06-01 17:24:29.188463-04:00 check_reload_status 507 Linkup starting re1This was that outage? For 4 seconds?
Do you have more that one gateway shown?
-
@stephenw10
Oh no, the WAN link was offline from 2026-06-01 15:57 until 2026-06-01 16:08. The power outage was shorter than that, although I am using logs from other devices to determine this.Only one gateway.
-
Hmm OK. So do you know what bounced the link on re1 at 17:24? I assume re1 is the WAN.